Combining 13C-NMR Triad Sequence Data with Joint Molecular Weight and Composition Data to Estimate Parameters in a Gas-Phase Polyethylene Reactor Model

Abstract

A three-site metallocene catalyst is used in a gas-phase semi-batch reactor to produce ethylene/hexene copolymers. At the end of each batch, polyethylene (PE) is collected and analyzed to determine the carbon-13 nuclear magnetic resonance (¹³C-NMR) triad sequence distribution. Joint molecular weight (MW) and composition distribution data are obtained using gel permeation chromatography with an infrared detector (GPC-IR). Data from ten experimental runs are used for kinetic parameter estimation. Using a mean-squared error (MSE) selection methodology, 23 of the 36 model parameters are selected for estimation using the available polymerization rate and PE characterization data. The remaining parameters are held at initial guesses to avoid overfitting. Addition of the triad data to the parameter estimation problem allows for one additional parameter to be estimated and results in improved parameter estimates. Standard deviations of all but one of the estimated parameters decreased due to inclusion of triad data. The updated parameter estimates result in good fits for the triad data and for joint MW and composition data. The model accurately predicts four validation data sets not used for parameter estimation. The new model and its updated parameter estimates will be valuable for scaling up new polymer grades from laboratory-scale to commercial-scale.